Comments on the Second Position Paper on the Evolution of ICE LIBOR

Transcription

1 October 16, 2015 Comments on the Second Position Paper on the Evolution of ICE LIBOR Japanese Bankers Association We, the Japanese Bankers Association ( JBA ), would like to express our gratitude for this opportunity to comment on the Second Position Paper on the Evolution of ICE LIBOR issued by the ICE Benchmark Administration ( IBA ) on July 31, We respectfully expect that the following comments will contribute to your further discussion. We would like to comment on the following 26 questions out of Questions 1 to 47 in the Second Position Paper. Q1: Do you agree in principle with using corporates as counterparty types? A1: YES We generally agree in principle with using corporates as counterparty types since we understand the significance of the requirements proposed in the report of the Financial Stability Board (FSB). However, full consideration needs to be paid to the implication arising from changing the original nature of LIBOR as a benchmark for interbank funding. Further, a change in the nature of LIBOR could have an impact on not only the interest rate level of LIBOR but also the pattern of rates movement. It is therefore also important to note that the predictability for the users may be affected. If the transactions with corporates as counterparty shall be included in the data to be used, it is desirable to include only transactions with large corporates in order to minimise the distortion from individual relation factors with each corporate clients. Further, if this suggested approach is going to be accepted, objective criteria for the distinction of corporate size will be also needed. 1

2 Q2: Do you think that transactions with Corporates as counterparty should be included with no premium or discount to adjust the transacted prices? A2: YES We do not assume that the proposed approach will suit the current concept and definition of LIBOR. At the same time, the proposed approach may be justified by an upcoming evolved definition of LIBOR and its evolved underlying interest. Premiums and discounts arising from a relation factor are considered to have little direct link to general market movements in terms of the interest rate. Consequently, these should not be included in the current LIBOR methodology. However, in concept, there may be an approach which includes premiums or discounts without adjustments. If such upcoming LIBOR methodology can be designed to dilute potential impacts from premiums and discounts with other (currently unidentified) factors, the impacts might be neglected. Q3: Do you think that the minimum size threshold should be increased for transactions with Corporates? If Yes, please explain your rationale and state what you think the increase in the threshold should be. A3: NO Even if transactional data with corporates is used for calculating submissions, we believe it is more appropriate to limit transactional data to with large Corporates whose market activities are comparable to banks and other financial institutions. We, however, do not think it is reasonable to increase the threshold of transactional data for the corporates to higher level than that for interbank transactions. Q4: Do you agree with IBA s proposal to maintain an Approved List of Funding Locations? A4: NO (Comment from the viewpoint of the LIBOR panel banks) Although this comment may conflict with a part of viewpoint of users, it is desirable to hire following process to determine the List of Funding Locations; firstly the benchmark submitters propose their own List of Funding Locations, then the IBA approves this, since the 2

3 funding structure may vary by currency and by panel bank. In other words, it is requested to carefully consider not to apply a uniform List of Funding Location, which includes transactions executed in funding centres outside London, to all panel banks. Inclusion of transactions executed in NY or Tokyo funding centres may deviate LIBOR submission from its underlying interest because there should be a significant gap in timing of data acquisition due to time differences. This is not desirable in terms of predictability of the benchmark. Q5: Do you agree that no weighting should be applied to adjust the transacted prices from different funding locations? A5: NO A careful consideration is required for including transactions executed in the NY or Tokyo funding centres. If these transactions are included, submission rates may deviate from the underlying interest of the benchmark because there is a significant gap in timing of data acquisition due to time differences. Accordingly, if data of transactions executed in funding centres outside London is going to be used, it is reasonable to apply a higher weight to London market data. Q6: Which of the four implementation options do you think is best? A6: Option 4 Option 4 is preferable. To maintain a framework to submit rates in accordance with the definition of LIBOR, which is the rate as of 11.00, is easier to understand in terms of the predictability of the benchmark. As long as the framework is maintained, even if an unexpected event has occurred in past 24 hours, it is expected that panel banks can provide their rate submission with a certain consistency. (Comment from the viewpoint of the LIBOR panel banks) Option 4 or 1 is preferable. As to Option 4, however, it is necessary to pay attention not to rely heavily on Expert Judgment, whereas, as to Option 1, underlying interest may not be representative if an unexpected event has occurred. 3

4 Q7: Would you find any of the implementation options problematic? A7: YES (Comment) Option 3 may have a significant impact on users since it entails a change in the publishing time of LIBOR. Q8: Do you think that LIBOR should be determined by reference to a point in time (e.g. as of London time) or by reference to a period of time (e.g. a 24 hour window)? A8: n/a (Comment) It is desirable to determine LIBOR by reference to a certain period of time, in order to collect actual transaction data to the greatest extent possible. Q11: If you think that LIBOR should be determined over a period of time, what period would be optimal in your view? Please state your preferred timing and explain your rationale if you have not done so in response to the questions above. A11: n/a In our view, one of the options would be 24 hours from previous submission. However, since the market may drastically change or an unexpected extraordinary event may occur during this period, it is necessary to be reminded that this period of time is not always the optimal option. Q13: Do you think that there would be merit in treating the Overnight tenors differently because of the different value date? A13: YES At least value date and settlement date of transactions on which the LIBOR Overnight tenor is based should be clearly specified. Q14: Do you agree with using FRNs / FCDs to supplement the Level 1 transaction types if necessary? 4

5 A14: NO Floating Rate transactions is not consistent with the interest period of the benchmark s tenors because such transactions are priced including a liquidity premium up to the maturity. It is therefore requested to clearly specify the logic of how these transaction data are treated in calculation. Q18: Which of the thresholds do you think is more appropriate? A18: USD 10 million and two trades Between the proposed two options, the threshold of USD 10 million and two trades is considered to be a better proposal because this option enables to collect more actual transactions comparing to the other. Q20: Do you think that a uniform threshold size should be applied across all currencies or that different thresholds should apply to different currencies/tenors? A20: NO (Comment from the viewpoint of the LIBOR panel banks) It is desirable to apply different thresholds to different currencies since the liquidity and the transaction volume differ by currency. On the other hand, a uniform threshold size should be applied to different tenors for each currency, since the submission practice may become too complicated if the threshold is changed according to the tenors. Q23: Do you think that different considerations should apply over month / quarter / year ends (when there is typically higher volatility in some currencies)? A23: YES It is beneficial to apply different considerations over month / quarter / year ends. This point relates to an issue of whether LIBOR continues to be a benchmark of interbank interest rates, or evolves to an interest rate benchmark that includes the wholesale market. On the other hand, due consideration should be made, instead of placing an importance on uniformity which limits all kinds of volatilities including those occurred from a sound reason, 5

6 such as a volatility observed as a result of seasonal factor in the money market. Q24: Do you agree with using transactions from previous day(s)? A24: YES We agree with the proposed approach given the basic policy to prioritize transactions is adopted. It should be noted that, however, data derived from historical transactions executed in the distant past could impair the benchmark s credibility as a representative of prevailing market as of the timing of rate submission. Q27: Do you agree with IBA s proposed decomposition formula for interpolation? A27: NO The proposed decomposition formula is complicated and not easy to understand for users. Q29: Do you agree that interpolation should not be applied to ON or 1W tenors? A29: YES The logic of applying interpolation to calculate 1W tenor is clear, whereas the logic of applying interpolation to ON could not be well understood. Q30: Do you think that interpolation with more than one tenor gap is acceptable? (e.g. if a bank has transactions for the 1M and 6M tenors, can the 2M and 3M tenors be interpolated?) A30: YES The proposed interpolation can be acceptable from a point of view which emphasizes automaticity of calculation. Q31: What are your views, in the absence of anchor points in the relevant currency, on interpolation from transactions in other currencies (e.g. EUR and USD for CHF and JPY) using FX swaps? 6

7 A31: n/a In our view, using FX swaps to calculate submission rates would be likely to contradict with the concept that the LIBOR is a benchmark of unsecured funding. Specifically, even in the case where FX swaps need to be considered as actual transaction data, as another option, Expert Judgment/Adjustment should also be considered, instead of simply using that data. Q32: Do you agree with the application of linear extrapolation or the decomposition formula? A32: n/a We do not agree with the proposed approach. If linear extrapolation is applied to 12M tenor, the resultant rate might significantly deviate from prevailing market rates. The decomposition formula is also complicated for users. Q33: Do you agree that extrapolation should not be applied to the ON or 1W tenors (because they are shorter, more liquid and more volatile)? A33: YES We agree that extrapolation should not be applied to the ON or 1W tenors because the extrapolation, in general, produces unnecessary volatility and undermines predictability. Q41: Do you agree in principle with having a concise description of LIBOR for users of the benchmark? A41: YES We believe that the disclosure of the definition is important to ensure transparency. Q42: Do you have any comments on IBA s proposed description of LIBOR for users of the benchmark? If Yes, please propose comments. A42: YES As users we have impression that, although the ICE LIBOR reform seeks a seamless 7

8 transition, the essential and definition of LIBOR would in fact change drastically. Accordingly, further explanation needs to be delivered to users about what reformed LIBOR will represent and that the seamless transition which IBA seeks will not have an impact on users. Q44: What would you see as the implications of changing the calculation methodology? A44: n/a Including data on actual transactions with corporates in the Level 1 could lead to a situation where the movement of LIBOR rates might be unpredictable by observing the movement in the money market. (For example, while the interest rate prevailing in the money market has declined from the day before, there will be cases where the LIBOR rate rises by effect of transactions with corporates.) Even though a seamless transition is proposed for LIBOR, it would be highly likely to become a benchmark with a different nature from the current LIBOR. Therefore, users shall need to consider whether proposed LIBOR could be used as the equivalent of the current one. In some cases, some users might need and reach an agreement on the use of the reformed LIBOR as the equivalent of the current LIBOR. Q46: Is there an alternative calculation methodology that you would prefer? Please describe the methodology and explain your rationale. A46: n/a If the essential of LIBOR is drastically changed as proposed, one option would be to simply develop a new benchmark, and discontinue the current LIBOR after a certain period of time. Q47: Do you agree that individual submissions should be published after three months on a non-attributed rather than on an attributed basis? A47: YES (Comment) We will agree with the proposal as long as comments from general users are collected and examined. 8